Industrial intelligence is increasingly based on agents, but how can a manufacturer ensure their agentic AI is not only reliable but also fast, open, and scalable, while remaining secure and robust? Velotic recently announced ThingWorx 10.2 industrial intelligence platform as the agentic AI backbone of Velotic’s AI strategy. AI is native to the platform and offered on existing customer subscriptions.
What’s New in ThingWorx 10.2?
To claim its spot as the backbone of Velotic’s agentic AI, this version has added an in-app AI assistant. Velotic discusses combining authentic human intelligence with artificial intelligence. This has broad implications not just for the Velotic software suite but also for the software that ThingWorx customers develop on this low-code platform. Manufacturers can now create custom agents and assistants with access to the company’s connected data sources.
ThingWorx is both a development platform and offers out-of-the-box (OOTB) applications for common and more complex needs. The Connected Work Cell application got an upgrade, and they added new apps for efficiency and service optimization as well. Velotic showed plans for pre-built agents for analysis, insights, configuration, integration, and development. Velotic is also keen to ensure customers provide guardrails, and they enable deterministic workflows.
Standards Support
Velotic has strengthened ThingWorx’s support for emerging standards to accelerate agentic success. These standards are for data handling, integration, and agent interaction.
MQTT: ThingWorx 10.2 has advanced MQTT support to improve data handling and modeling. Those pursuing a unified namespace (UNS) or a similar architecture know that MQTT can support OT data movement.
CESMII: This version also supports both Smart Manufacturing Profiles and the i3X API, a CESMII standard that focuses on machine-level interoperability and collaboration in multi-vendor environments. This partnership was announced in August.
MCP: This version supports the Model Context Protocol (MCP), which enables agents to interact with each other and to connect with external data sources and systems. MCP is a standard that is accelerating the orchestration of agents into more powerful and trustworthy systems. ThingWorx 10.2 can also act as an enterprise MCP server, enabling access to its rich data from external agents.
Rich Intelligence History
ThingWorx was one of the original industrial internet of things (IIoT) platforms, originating in 2009. The IIoT Platform category was intended to ensure data from machines and other “things” could deliver value in real-time industrial environments. With a strong model at its core, ThingWorx enabled rapid application development and user interface (UI) mashups, allowing non-developers to create useful visualizations.
Some ThingWorx customers have been connecting OT and IT data sources and building custom applications for their operations for a decade or more.
In 2025, as part of PTC, the ThingWorx platform added an AI-based chat function and predictive capabilities. Some of what is in 10.2 also came out in Beta last year: agentic AI and the MCP server. Today, Velotic claims ThingWorx has over 3,300 customers, who are evolving with it from IIoT to AI-driven intelligence.
Early User Case
One HVAC manufacturer has built a ThingWorx app that provides real-time visibility into OEE and machine degradation or downtime via a conversational agent. This company used the mashup builder and composer to create machine monitoring at five of their 27 sites. It is also connected to other plant and enterprise applications. This is improving uptime, shift handovers, and root cause analysis.
Building Velotic’s AI
As a newly formed company, Velotic is crafting its story right in the midst of the AI revolution. With 1200 employees, it has an aggressive roadmap for AI both within ThingWorx and across the portfolio, which also includes Proficy MES, SCADA, Historian, and Kepware. We expect to see agents and AI assistants across the portfolio in a year.
Our Take
We are excited to see the renewed investment in ThingWorx under Velotic and its position as the backbone for agentic AI. It makes sense to us, and we have had a chance to talk to happy ThingWorx customers over the years. Both the OOTB apps the company provides and those the customers build have already delivered structured value. That strong model or structure at the core of this platform could dispel many of the concerns and problems companies have had with industrial AI efforts to date. A strong understanding of both manufacturing and service puts the ThingWorx team in a position to continue supporting customers throughout the product lifecycle.
Thank You
We appreciate being part of the Velotic trusted analyst community on this briefing. Thank you to Ayush Tiwari for providing the product information, and to Carver Conway for keeping us in the loop.